OpenAI Lawyers Allege Shivon Zilis Served as Elon Musk's Covert Liaison Inside the AI Lab

# OpenAI Lawyers Allege Shivon Zilis Served as Elon Musk's Covert Liaison Inside the AI Lab

A February 2018 text message, now entered into evidence in the blockbuster Musk v. Altman trial in Oakland, may be the most revealing sentence produced so far in a case already overflowing with them. Shivon Zilis, a Neuralink executive and mother of four of Elon Musk's children, wrote to the billionaire: "Do you prefer I stay close and friendly to OpenAI to keep info flowing or begin to disassociate?"

That single line, dated February 16, 2018, has become a centerpiece of OpenAI's defense in a trial where roughly $130 billion in value hangs in the balance. OpenAI's legal team argues the message is a smoking gun -- proof that Zilis functioned not as an independent adviser and board director, but as Musk's covert eyes and ears inside the artificial intelligence lab he cofounded and then left.

A Dual Role Spanning Seven Years

Zilis's entanglement with both Musk and OpenAI runs deep. She joined OpenAI as an adviser in 2016, the same year she and Musk began a romantic relationship. She was elevated to the nonprofit's board of directors in 2020, a seat she held until 2023. Throughout that period she also worked at Neuralink, Musk's brain-computer interface company, where she held the title of director of operations and special projects.

The timeline matters because Musk departed OpenAI's board in February 2018 -- the very month Zilis sent the "keep info flowing" text. OpenAI's lawyers, led by William Savitt of Wachtell, Lipton, Rosen and Katz, contend that even after Musk formally exited, Zilis continued channeling strategic intelligence back to him, including details about personnel, research direction, and the organization's evolving corporate structure.

An email from August 28, 2017 -- months before Musk's departure -- underscores the pattern. In it, Zilis informed Musk's chief of staff Sam Teller that she had met with OpenAI president Greg Brockman and cofounder Ilya Sutskever to discuss how equity would be distributed in a proposed restructuring. She outlined two paths: "Roll everything into a B corp" or create a separate "OpenAI C Corp and OpenAI non-profit." That restructuring would eventually become the 2019 for-profit conversion at the heart of the lawsuit.

Musk's Testimony: Selective Memory

When Musk took the stand earlier this week, he was asked to describe his relationship with Zilis. "We live together," he said. "She's the mother of four of my children." He clarified that her role was "senior adviser" rather than chief of staff, a distinction that appeared aimed at minimizing the perception of operational coordination.

Pressed by Savitt on whether Zilis ever shared confidential OpenAI business information with him, Musk said he did not recall. The answer drew visible skepticism from the jury box, according to courtroom reporters, particularly given the documentary evidence already in the record.

Savitt also confronted Musk with allegations that he had instructed Zilis to file paperwork to convert OpenAI into a for-profit entity -- a move that would have given Musk a majority equity stake. Musk again said he did not recall. OpenAI's lawyers argue that the failed 2017 negotiations collapsed precisely because Musk demanded "absolute control" of the restructured entity, a demand Brockman and Sutskever refused.

The Broader Trial Context

The Zilis subplot sits within the larger Musk v. Altman case, which opened April 28 in Alameda County Superior Court before Judge Evelio Grillo. Musk is seeking up to $150 billion in damages, alleging that OpenAI's transformation from a nonprofit research lab into a for-profit juggernaut constituted unjust enrichment and a breach of charitable trust. His damages expert, C. Paul Wazzan, testified that the nonprofit foundation's current 26 percent stake in the for-profit arm is worth approximately $130 billion.

OpenAI's defense strategy has been to recast Musk not as a betrayed philanthropist but as a spurned power broker -- someone who wanted controlling interest in what has become the most valuable AI company on earth, and who launched a rival venture, xAI, when he could not get it. The Zilis evidence fits neatly into that narrative: if Musk maintained an intelligence pipeline into OpenAI even after leaving, it suggests his interest was strategic rather than charitable.

Analysis: Why the Zilis Evidence Could Be Decisive

Corporate governance cases often hinge on questions of loyalty and information asymmetry. If the advisory jury concludes that Zilis was effectively a Musk agent during her tenure on the OpenAI board, the implications extend beyond this trial. It would raise questions about whether her board votes -- including decisions about the for-profit conversion, Microsoft's multibillion-dollar investment, and research priorities -- were made in OpenAI's interest or Musk's.

Zilis is expected to testify later in the trial, alongside Microsoft CEO Satya Nadella, former OpenAI CTO Mira Murati, and Ilya Sutskever. Her testimony could either defuse or detonate the liaison narrative. If she acknowledges acting as an intermediary, it strengthens OpenAI's portrayal of Musk as a controlling figure who never truly let go. If she denies it, the text messages and emails will speak for themselves.

The trial continues next week. Whatever the outcome, the Zilis revelations have already reshaped the public understanding of how power, romance, and artificial intelligence intersected inside one of the most consequential technology organizations of the twenty-first century.